How the scan works
Siteiz measures one thing: how easily AI systems, the crawlers and assistants that increasingly read the web for people, can understand your page. Here is the approach in plain terms, and what the score does and does not tell you.
We read your page twice
We fetch your page the way an AI crawler does, as plain code with no JavaScript run, and the way a person does, in a real browser that loads everything. Then we compare the two. The difference between what a visitor sees and what a machine receives is where most sites quietly lose visibility in AI search, because many AI crawlers do not run the JavaScript your site depends on.
What we look at
Every scan grades eight areas that decide whether AI search can reach, read, understand, and quote your page. The percentage is how much each area counts towards the grade:
- Crawl access (15%). Whether AI crawlers are allowed in, and whether anything quietly blocks or de-indexes the page.
- Visibility (20%). How much of what people see is in the code a crawler receives, without running JavaScript.
- Structure (15%).One clear H1, a clean heading outline, and the page's own content standing out from menus and footers.
- Machine facts (15%).Structured data that describes what the page is about, who publishes it and where it sits in the site, plus the title, description, canonical URL and OpenGraph tags. Generic types earn little; the page's subject earns the most.
- International (10%). Whether the page declares the language it is written in, and links its other language versions with valid hreflang.
- Actionability (10%). Whether an agent can follow your links, navigation, and calls to action, not just a human with a mouse.
- Citability (10%). Concrete facts an AI answer can quote, and a date that shows the page is current.
- Efficiency (5%). How much code a machine has to read per useful word, since bloated pages get cut off.
Each area is scored from 0 to 100 and combined into one grade from A to F. Most areas hold whether or not a crawler runs JavaScript, so the score stays meaningful as crawlers improve.
Do AI bots get the same page?
Some sites and CDN tools serve AI crawlers a different version of a page than people see. We check for that separately. We request your page twice as a browser and once each as GPTBot, ClaudeBot, and PerplexityBot, ignore text that changes on every request, and compare the rest. You see the exact text, title, description, and structured data that only the bots receive, or that they miss. If your server offers a markdown version, we compare that too. A lighter format is fine. Different content is the cloaking risk.
This check is reported as evidence and does not change your grade. Some sites verify real crawlers by IP address and refuse our test request. We report that as unverified, never as a match.
Evidence, not opinion
Every result is measured from your actual page and shown with the evidence behind it, the real titles, tags, links, and facts we found, so you can see exactly why each point moved. The scoring is consistent: the same page always produces the same score. There is no AI guessing a grade, and we never invent numbers. If a page blocks our reader or gives us too little to work with, we say so and withhold the grade rather than dress up an empty result.
What the score does not claim
A scan looks at one page at one moment, not your whole site. It measures how machine-readable your page is, not whether your content is good, true, or certain to be cited. Some AI crawlers are beginning to run JavaScript, so a gap we flag is a real risk and a real cost, not a guarantee of invisibility. We are deliberate about these limits, because a tool you can trust is worth more than one that overpromises.
Want to see it on your own site? Run a free scan.